#graphsearch

What will facebook's graph search expose?

Facebook started trialling graphsearch in January 2013 with a limited beta rollout.   With the help…

"Graph Search is a powerful recruiting tool." a déclaré Mark Zuckerberg. Cet outil qui ne laisse personnes indifférent s'appreterait-il à bouleverser le social recruitment ?

Nul doute que Facebook et la mine d'informations qu'il possède peut être l'outil de recrutement dont rêvent les RH. A quelques mois de la sortie du Graph Search, beaucoup annoncent Facebook comme étant en passe de devenir la plus grande cvthèque du monde.

Ainsi, grâce au Graph Search, les recruteurs pourront rechercher parmi leur réseau qui travaille dans telle entreprise, à quel poste, depuis combien de temps. Certes, ces informations restent encore très incomplètes pour procéder à un recrutement, mais avec le croisement de données et les utilisations multiples des adresses mail, on imagine mal Facebook se priver  d'un accès aux données posséedées par d'autres réseaux sociaux.

Lance Haun a pu tester le Graph Search pour le comparer avec le moteur de recherche de membres de Linkedin, et il est apparu que Linkedin donnait de bien meilleurs résultats.

The amount of digital data available to us is increasing rapidly. Cheap storage space, increasing internet speed and distributed computing has enabled everything to be done and hosted online in cloud. We are in the age of Big Data, where almost anything – information from every sensors, every entries of every weblogs and tweets, every search queries typed in – could be stored easily. To navigate through as much as petabytes or even exabytes of data, we rely heavily on search engines. As the digital data keep growing, more improvements are made to the search engines as well, but speed is not the only change we will see. Recent updates to technologies surrounding searching information seem to hint that we are going toward different types of searches.

Facebook Graph Search In January, Facebook started beta testing the new search in its social networking service. Graph Search tries to optimize the results to best reflect the likings of you and your close friends. The search is based on information of who are friends, who liked what, and all other personal preferences and interests users have disclosed on the facebook. Using these information, it attempts to personalize your search results – so when you search for restaurants, ideally it will recommend a restaurant your friends liked so you have something to talk with your friends. In a sense, it has characteristic of recommendation an e-commerce website gives you. Graph Search will not function well as a regular search engine to look for information, but it will be a powerful tool to help you socialize.

Hashtags When we search by tags, it does not make much difference user-experience-wise as searching by keywords like in regular searches. We just type into the search box what we are interested in, and the results return relevant media that includes the words we have typed in. The advantage of hashtags lie in bringing together unorganized contents and creating connection in between. When you use hashtags, you know how to join conversation and share your opinion about the particular topic you are tagging. If anyone gets interested, they could immediately click the tag to see what other people are saying. No matter how diverse the actual opinions might be, you can “connect” with others in not just pre-defined but any random topic you choose. Just like graph search, hashtag helps you explore your interests and possibly connect with other people. It lets you create your own list of tags you can casually click on to search and form loose communities around particular topic.

Neural Networks On March 12, Google bought up a startup company started by Professor Geoffrey Hinton at University of Toronto and two of his students. This company, DNNresearch, centered on research of neural networking, which might bring new powers to search engines. Neural networks is a mathematical model that tries to simulate human brain mechanism, which in physical sense is interaction of neurons turning on and off. Professor Hinton's research already has implications in areas such as speech recognition and computer vision. Search engines have traditionally focused on searching texts but not images or audios because images and audios cannot be compared easily. Human brains could look at objects from different angles and recognize them as same object, or listen to speeches in different accents and still understand they are same language, but these tasks have been very difficult for computers to achieve. Neural networks models human brain and essentially will try to simulate human brain in a computer. The advance in this research could possibly mean a future where we could search for songs by melody, look up movies using screenshot, or search and jump to particular topic in a long lecture recording.

additional resources:

In an effort to keep its user hooked to the Social Networking site Facebook has launched a revamped News Feed. Currently a Facebook user spends on average 22 minutes online.

The new design that incorporates bigger images, allows customization and may attract advertisers in order to increase the marketing on Facebook. Chief Executive of Facebook Mark Zuckerberg referred to the new design as a "Personalized Newspaper".

All Facebook users are to receive this new feature before Facebook celebrates its 10th birthday next February.

Photos and video will now have more space on the News Feed, which has changed Facebook’s content in the past few months. Approx.  50 percent of the posts on News Feed is either a photo or a video.

 This change is likely to influence Facebook advertisers to be more creative in their marketing pitches as rich media gains most attention across all the Social Media platforms.